Short answers, pulled from the story.
The Turing test is a method for assessing whether a machine can exhibit intelligent behaviour indistinguishable from a human. A human evaluator reads a text transcript of a conversation between a human and a machine and tries to identify which is which; if the evaluator cannot reliably tell them apart, the machine passes. Alan Turing introduced the test in his 1950 paper "Computing Machinery and Intelligence."
The original imitation game was a party game involving three players: a man, a woman, and an interrogator. The interrogator, unable to see the others, sent written questions and tried to determine which player was the man and which was the woman. Turing's 1950 paper replaced the man with a computer and asked whether the interrogator would be fooled just as often.
In March 2024, Stanford researchers reported that ChatGPT-4 passes a rigorous Turing test, diverging from average human behaviour chiefly by being more cooperative. A late March 2025 study found that GPT-4.5, when instructed to adopt a humanlike persona, was identified as the human 73 percent of the time, significantly more often than actual human participants.
In his 1980 paper Minds, Brains, and Programs, John Searle argued that a program can pass the Turing test purely by manipulating symbols it does not understand. Because the program lacks understanding, Searle concluded it cannot be said to think in the same sense humans do, and the test therefore cannot determine whether a machine truly thinks.
The Loebner Prize was an annual competition for practical Turing tests, first held in November 1991 and underwritten by Hugh Loebner. The bronze medal for most human-seeming conversational behaviour was awarded each year; A.L.I.C.E. won it in 2000, 2001, and 2004, while Jabberwacky won in 2005 and 2006. The competition ended in 2019 after Loebner's death in 2016.
CAPTCHA stands for Completely Automated Public Turing test to tell Computers and Humans Apart and is a form of reverse Turing test, in which a machine challenges a user to prove they are human. Original CAPTCHA versions displayed distorted letters or images for humans to identify; reCAPTCHA v3 runs invisibly in the background. By 2014, Google engineers had demonstrated a system that defeated CAPTCHA challenges with 99.8 percent accuracy.